This document describes the data files used for all analyses in ‘Veltsos et al. Sexual selection on song and cuticular hydrocarbons in two distinct populations of Drosophila montana’.
There are 3 files.
MaleSong.txt- data on males and their song used in song only analysis (526 individuals).
FemaleCHC.txt- data on females and their CHCs used in the female CHC analysis (457 individuals).
MaleSongCHC.txt- data on males and their song and CHCs, used in the combined analysis (442 individuals).
The first columns in the text files were used in the models. They were 14, 20 and 14 variables for
MaleSongCHC.txt respectively (up to ‘timeSong’ in all datasets). The remaining columns contain raw data that were used to generate some of those first columns (for example the CHC data from which principal components were calculated).
There is a smaller lower number of individuals in
MaleSongCHC.txt compared to
MaleSong.txt because some males were lost before extracting CHCs, or because CHC extraction data were of low quality. There were no individuals with CHC data but no song data.
gen - Generation: The only random variable in any model.
pop - Population: Vancouver or Oulanka.
temp - Temperature.
timeCopul - Time to mating since song started, which is a proxy for when the flies start interacting. It was used in the copulation latency models.
mating - Binomial trait, used to estimate the significance of explanatory variables for mating success within 10 min.
rel_mating - Mating from above, converted to relative mating for estimating effect sizes for mating success within 10 min.
resist and rel_resist - As for mating, for the models with female rejection song.
song(PN, FRE, IPI, CN)_norm - Normalised song characters so that their variation is comparable.
song_norm_sq - Square term for the same values, added in models to estimate quadratic effects.
CHC_MF_pc# - (note, # stands for a number). CHC principal component #, used in the models. CHC proportions to total CHC in an individual (e.g. C#.#pc/di) were log transformed (to C#.#pc/di.logcontr) to generate principal components using data from both sexes (notNormCHC_MF_pc#). These were transformed to have a total variance of 1 (by dividing with the square root of the sum of the variance of all the scores) to make the song and CHC PCs directly comparable.
CHC_MF_pc#_sq - The above numbers squared, they were used as explanatory variables to estimate quadratic selection on CHCs.
song_pc# - Principal components of the normalised song characters (song_norm), after being transformed to have a total variance of 1.
song_pc#_sq - The above numbers squared, to estimate quadratic selection in the models.